Abstract

Higher than acceptable aflatoxin levels in peanut kernels (Arachis hypogaea L.) and related products is a worldwide food safety concern. Strict regulatory standards by major importers of peanuts limit the marketability of peanuts for many developing tropical countries including Zambia. The incidence of pre-harvest aflatoxins is strongly linked to soil and weather conditions during pod-development. This study aimed to formulate statistical models to predict total aflatoxin content in peanut kernels using selected environmental factors during pod development. Field experiments were conducted for two years during which the peanut crop was exposed to 84 combinations of ambient temperature, soil temperature and soil moisture content measured during the last 30 days of pod development. These data were used to formulate regression models to predict total aflatoxin content in peanut kernels. Simple linear regression models had R² values of 0.30 for maximum ambient temperature, 0.24 for soil temperature and 0.38 for soil moisture content. Combining soil moisture content and soil temperature in a multivariate regression model could explain 54% of the variation in total aflatoxin content while a combination of soil moisture content and maximum ambient temperature could only explain 46% of the variation in total aflatoxin content.

Key words: Aflatoxin, groundnut, linear regression, statistical model, Zambia.